Metraipolis › archive › dc_ea1f3cdd7403
A Latent Distribution Perspective on Evaluating and Improving Latent Generative Models — arXiv:2609.24088
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.24088
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T22:04:59.938Z
- SHA-256 of the archived text
e42629020620b06df5efe5d56cd3f11b809be9cc7e0d3512e3dd92353806b5ce- Stable URL of this page
https://metraipolis.com/documento/dc_ea1f3cdd7403
How to cite
A Latent Distribution Perspective on Evaluating and Improving Latent Generative Models — arXiv:2609.24088 — arXiv (en), https://arxiv.org/abs/2609.24088 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ea1f3cdd7403 (Metraipolis archive dc_ea1f3cdd7403, sha256 e42629020620b06df5efe5d56cd3f11b809be9cc7e0d3512e3dd92353806b5ce)
BibTeX
@misc{dcea1f3cdd7403,
title = {A Latent Distribution Perspective on Evaluating and Improving Latent Generative Models — arXiv:2609.24088},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.24088},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ea1f3cdd7403; sha256 e42629020620b06df5efe5d56cd3f11b809be9cc7e0d3512e3dd92353806b5ce},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24088. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
In latent generative models, reconstruction quality is often assumed to correlate with generative performance. However, reconstruction FID (rFID) can exhibit weak or even negative correlation with generation FID (gFID). We attribute this misalignment to a latent distribution mismatch: reconstruction evaluates the decoder on encoder-induced latents, whereas generation uses the same decoder on latents produced by the generative model. To characterize this shift, we introduce generation-aware reconstruction (GAR), which constructs a continuous trajectory from standard reconstruction toward generation by perturbing encoder latents with noise and denoising them through the generative model before decoding. GAR probes the decoder behavior along this trajectory, making the transition from encoder to generation-time latent distributions observable and diagnosable. The resulting trajectory-based diagnostic, GAR-FID, exhibits strong empirical correlation with gFID across diverse tokenizers and scales. Importantly, intermediate GAR latents become more generation-aware while preserving correspondence with their source images, thereby retaining paired supervision that is absent for fully generated latents. This correspondence enables decoder adaptation on intermediate GAR latents, consistently improving generative quality across model scales. Overall, latent distribution mismatch provides a useful perspective for evaluating and improving latent generative models. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.24088 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-09T21:45:20.980Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).